Refactor code structure for improved readability and maintainability; optimize performance in key functions.

scripts/corpus/validate-corpus.py

@@ -0,0 +1,54 @@
+#!/usr/bin/env python3
+"""Validate golden corpus case structure."""
+from __future__ import annotations
+
+import json
+import sys
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[2]
+CORPUS = ROOT / "bench" / "golden-corpus" / "categories"
+
+REQUIRED = [
+    "case-manifest.json",
+    "run-manifest.json",
+    "input/sbom-cyclonedx.json",
+    "input/sbom-spdx.json",
+    "input/image.tar.gz",
+    "expected/verdict.json",
+    "expected/evidence-index.json",
+    "expected/unknowns.json",
+    "expected/delta-verdict.json",
+]
+
+
+def validate_case(case_dir: Path) -> list[str]:
+    missing = []
+    for rel in REQUIRED:
+        if not (case_dir / rel).exists():
+            missing.append(rel)
+    return missing
+
+
+def main() -> int:
+    if not CORPUS.exists():
+        print(f"Corpus path not found: {CORPUS}")
+        return 1
+
+    errors = []
+    cases = sorted([p for p in CORPUS.rglob("*") if p.is_dir() and (p / "case-manifest.json").exists()])
+    for case in cases:
+        missing = validate_case(case)
+        if missing:
+            errors.append({"case": str(case.relative_to(ROOT)), "missing": missing})
+
+    if errors:
+        print(json.dumps({"status": "fail", "errors": errors}, indent=2))
+        return 1
+
+    print(json.dumps({"status": "ok", "cases": len(cases)}, indent=2))
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())